Binyu

Product Manager building AI memory @Memoket_AI
Former PM @Tencent · former banker @CreditSuisse
Building boldly. Living fully.

[Binyu @Binyu63]

My biggest takeaway as a product manager:

Marketing is not only user acquisition.
Marketing is continuous product discovery.

Every public narrative is a reflection of how we understand our users.
Every user response refines how we build the product.

All growth eventually circles back to one thing:
Real product value creates authentic stories, not the other way around.

Storytelling communicates value.
Distribution amplifies value.
Iteration upgrades value.

Sustainable growth happens when the three reinforce each other endlessly.

What’s your take on narrative vs product-market fit?
I’d love to exchange thoughts with builders & PMs here.

[Binyu @Binyu63]

Some recent deep product reflections that reshaped how I understand early-stage growth:

Storytelling, product iteration, distribution, and PMF are not four separate functions.

They are different parts of one single, continuous process.

Most importantly:
Storytelling is never just about “marketing our product.”
It’s how we discover who truly needs it, why they need it, and how to reach them.

[Binyu @Binyu63]

Conversion data is how the market talks back to product teams.

Views and likes are vanity metrics.
What truly matters is the quality of users we attract:
• Who actually reads and clicks through?
• Who signs up and tries the product?
• Who retains, pays, and refers?

Different narratives attract completely different user groups.

That’s why we need three layers of fit:
Narrative-Market Fit — who feels understood by our story
Channel-Market Fit — where our ideal users live
Product-Market Fit — whether we deliver lasting value

They iterate together, not in sequence.

[Binyu @Binyu63]

Memoket Desktop Is Here with 22 Work Apps https://t.co/6merm94jJP

[Binyu @Binyu63]

A correct edit in the wrong place is still a wrong edit.

Our engineer built a safeguard that rejects ambiguous replacements instead of guessing which passage to change.

A small detail that makes it easier to trust AI with your documents.

[Kris Tan @memokris]

Put the same sentence in two paragraphs, then ask for a replacement. A first-match search can make a perfectly reasonable edit in the wrong place. Our revision check rejects an ambiguous replacement unless the target span is bounded.

[Binyu @Binyu63]

One thing I’ve been questioning recently is whether we’re thinking about “memory” the right way.

A good summary helps you understand what happened in one conversation.

But most things we actually care about don’t happen in a single conversation.

A project evolves over weeks.
A customer changes their mind.
The same issue comes up three times.
A decision gets revised.
Something mentioned a month ago suddenly becomes important again.

So my current intuition is:

AI memory shouldn’t just help us remember more.

It should help us understand what from the past matters now.

For example:
- What changed since the last conversation?
- What keeps coming up?
- What’s still unresolved?
- Did something we discussed weeks ago suddenly become relevant again?
- What should I remember before the next conversation?

The interesting part to me is that none of this really lives inside a single summary.

It lives between conversations, over time.

We’ve also been exploring the technical side of long-term memory at Memoket. If you’re interested in that side of the problem, we have an open-source project called KITE on GitHub as well.

But right now I want to test the product hypothesis with real people and real conversations before deciding what this experience should actually become.

So I’m looking for a few people who have one ongoing topic, project, customer, or decision that has appeared across multiple recorded conversations.

For the experiment, I’ll ask you to choose around 4–6 conversations from different days or weeks.

The transcript is the main input. Audio is only helpful when I need to verify context.

I’ll use those conversations to prototype what a cross-conversation memory experience could look like, and I’ll share the result back with you.

You choose exactly what you’re comfortable sharing, and I’ll only use the conversations you intentionally provide for this experiment.

Please only share conversations you’re comfortable and permitted to share.

If this sounds useful and you’d like to try it, DM me.

[Binyu @Binyu63]

When a topic evolves across multiple conversations, can AI help users recover and understand the past context that matters now — in a way that a single summary or search cannot?

[Binyu @Binyu63]

@Memoket_AI The four integrations are listed in Memoket for iOS, v1.2.10. If you want to try them: https://t.co/4SKs2K45KK

[Binyu @Binyu63]

New at @Memoket_AI: Discord, Google Drive, OneDrive and Confluence integrations.

I almost posted a grid of logos. But the harder question is: what happens to a decision as it moves from a Discord thread to a file to a team doc?

Where does context get lost for you?

[Binyu @Binyu63]

Onboarding non-tech users and a little touch make a huge difference.

[Emil Kowalski @emilkowalski]

Interesting to see how the AI companies are trying to make their apps feel more human, less alien.

Muse has this cute little animated guy smiling all the time, ready to help you with anything. It feels friendlier and more relatable than a blank, sterile screen with an input.

ChatGPT Computer Use shows a moving cursor as it’s working. It’s not technically required, but it makes the whole process less scary and easier to follow.

It’s like the early iPhone days when things used to immitate real world objects so that an unfamiliar technology felt more intuitive and relatable.

In this case we imitate presence, trying to give the AI a more human feel.

[Binyu @Binyu63]

@CreatorAstrojan 这个我真的会看

[Binyu @Binyu63]

アップグレードの手続きについて混乱を招いてしまい、申し訳ありません。🙏
Pre-orderユーザーの方は、現在お持ちのPro特典を無駄にすることなくUnlimitedへアップグレードできますので、ご安心ください。
詳しい手続きについては、すでに [email protected] からメールをお送りしています。ご確認いただき、ご不明な点があればそのメールに直接ご返信ください。

[Binyu @Binyu63]

Pretty special seeing something we’ve been building every day show up at IFA like this 🥹
Still a lot to improve, but proud of how far Memoket has come.

[Memoket @Memoket_AI]

We're an IFA Innovation Award honoree 🏆

Memoket Gem's European debut is off to a strong start. Huge thanks to everyone who stopped by.

If you're at IFA, come chat with us at 5.2, Booth 452.

More coming this week 👀
#IFA2026 #IFABerlin #Memoket

[Binyu @Binyu63]

X上真的全是搞钱的信息差，欢迎大家多多交流！

[金尘马 @jinchenma_ai]

我操，太牛逼了。

X 上天天有新东西刷新我的认知。

链接上一个大哥，北美某 AI 创业公司 CTO。

大半夜的几句话，又透漏一堆赚钱的路子。

比如，OPC 出海，可以搞搞 Google Map 的 Site，就是帮小商铺老板装修一下他们在谷歌地图上的店铺，很多小老板不会搞。

类似于淘宝商城。

至于渠道，就是批量 cold mail，cold call，用 Pine 可以做。

或者找 nextdoor，类似国内的 58 同城。

妥妥信息差啊。

[Binyu @Binyu63]

Cant wait to try Astra today!!

[Tibo @thsottiaux]

We will give one banked reset for every day you don't have access to Astra on your paid ChatGPT plan, starting today. Team is moving mountains to give access as fast as we can.

First one will land in ~ 3 hours. There is still time to create your account if you don't have one.

[Binyu @Binyu63]

Hi Nicole — this sounds amazing! I’m also a product manager working on an AI product, and we’re still searching for PMF, so hypothesis testing is a big part of my work.

If this works well, it could save us so much time by helping us test hypotheses without having to ship every idea first. I’m going to try the app tomorrow and will get back to you with my thoughts!

[Binyu @Binyu63]

Been thinking about this a lot lately.

The real value of recording may not be better summaries, but turning conversations into context that agents can actually act on

[Zara Zhang @zarazhangrui]

We are not recording meetings for humans anymore. Nobody reads the AI summary notes or listens back. We are capturing transcripts to feed agents.